# VLA评测揭示意图与执行鸿沟：感知规划强但动作泛化失效

> 原文链接：https://robotics.hellobike.com/papers/2025-06-11-2506-09930-cd4fc318
> 频道：papers | 更新日期：2025-06-11 | 标签：auto, arxiv, cs.RO
> 核心摘要：研究发布包含50个仿真任务的统一评测套件，系统评估主流VLA模型。结果显示VLM骨干赋予模型良好意图与规划能力，但在分布外观测下动作执行显著失效，且动作微调会侵蚀原始推理能力，代码已开源。

# VLA评测揭示意图与执行鸿沟：感知规划强但动作泛化失效

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[From Intention to Execution: Probing the Generalization Boundaries of Vision-Language-Action Models](https://arxiv.org/abs/2506.09930)（arXiv 2506.09930）

## 核心内容

针对视觉-语言-动作（VLA）模型缺乏标准化评估及复现难的问题，研究团队推出包含50个仿真任务、覆盖10个子类别的统一探测套件，涵盖语言指令、视觉与物体操作维度。该工作系统评估了多款SOTA VLA架构的泛化边界，并开源完整任务集与评估代码，旨在建立行业基准以推动感知到动作闭环的研究。

评测结果揭示了VLA模型的“意图-执行”断层：尽管VLM预训练赋予了鲁棒的感知理解与高层规划能力（即良好的意图），但在面对分布外（OOD）观测时，策略虽能保持逻辑连贯却频繁在动作执行层面失败。此外，研究发现针对动作数据的微调过程存在副作用，会导致模型丧失部分原始VLM的通用推理能力，表明当前从语义理解到精准运动控制的迁移仍面临本质挑战。

## 关键看点

- **统一评测基准**：开源包含50个仿真任务及10个子类别的探测套件，解决现有VLA评估任务单一及真机复现壁垒问题
- **意图执行断层**：VLA在分布外观测下表现出连贯的高层规划意图，但底层电机动作执行精度显著下降甚至失效
- **微调能力侵蚀**：实验证实对动作数据的微调会削弱VLM骨干原有的通用推理能力，揭示感知与动作对齐的权衡难题
- **泛化边界探测**：系统性验证VLM预训练对下游机器人策略的贡献度，明确当前模型仅在感知规划侧具备有效泛化

## 工程与产业观察

该研究量化了具身大模型“懂意图但手笨”的工程痛点，警示业界单纯堆叠VLM参数无法直接解决运动控制泛化问题。开源的标准化仿真评测套件为算法团队提供了验证动作策略鲁棒性的关键基础设施，提示未来研发需重点关注动作微调中的灾难性遗忘抑制及Sim-to-Real的动作空间对齐技术。

## 英文摘要 (Abstract)

One promise that Vision-Language-Action (VLA) models hold over traditional imitation learning for robotics is to leverage the broad generalization capabilities of large Vision-Language Models (VLMs) to produce versatile, "generalist" robot policies. However, current evaluations of VLAs remain insufficient. Traditional imitation learning benchmarks are unsuitable due to the lack of language instructions. Emerging benchmarks for VLAs that incorporate language often come with limited evaluation tasks and do not intend to investigate how much VLM pretraining truly contributes to the generalization capabilities of the downstream robotic policy. Meanwhile, much research relies on real-world robot setups designed in isolation by different institutions, which creates a barrier for reproducibility and accessibility. To address this gap, we introduce a unified probing suite of 50 simulation-based tasks across 10 subcategories spanning language instruction, vision, and objects. We systematically evaluate several state-of-the-art VLA architectures on this suite to understand their generalization capability. Our results show that while VLM backbones endow VLAs with robust perceptual understanding and high level planning, which we refer to as good intentions, this does not reliably translate into precise motor execution: when faced with out-of-distribution observations, policies often exhibit coherent intentions, but falter in action execution. Moreover, finetuning on action data can erode the original VLM's generalist reasoning abilities. We release our task suite and evaluation code to serve as a standardized benchmark for future VLAs and to drive research on closing the perception-to-action gap. More information, including the source code, can be found at https://ai4ce.github.io/INT-ACT/

作者：Irving Fang, Juexiao Zhang, Shengbang Tong, Chen Feng
:::

## 原始参考来源 (Sources)

- [arXiv 2506.09930：From Intention to Execution: Probing the Generalization Boundaries of Vision-Language-Action Models](https://arxiv.org/abs/2506.09930)